Okay.
Okay, perfect. All right, good afternoon and welcome to the first day of JPMorgan's 53rd Annual Technology Media and Communications Conference. My name is Harlan Suhr. I'm the semiconductor and semiconductor capital equipment analyst for the firm. Very pleased to have Irving Tan, CEO of Western Digital. We have Kris Sennesael, the newly appointed Chief Financial Officer of the team. And here in the front, we have Ambrish Srivastava, Head of IR, here with us as well. Kris is going to kick us off with a safe harbor statement, and then we'll go ahead and kick off the Q&A. Thank you, gentlemen, for joining us today.
Yeah, Harlan, thanks for having us here at this great event. Just to start, I want to read this safe harbor statement. Today we are going to make some forward-looking statements and discussions based on management current assumptions and expectations, including with respect to our product portfolio, business plans, and performance, market trends and dynamics, and future financial results. These forward-looking statements are subject to risks and uncertainties. Please refer to our most recent financial report on Form 10-K and our filings with the SEC for more information on risks and uncertainties. We will also be making some references to non-GAAP financials, and our reconciliation of our GAAP and non-GAAP results can be found in our investor relations sections on our website. With that, I'll turn it back to Harlan.
Oh, great. Thank you. Thank you, gentlemen, for joining us again today. Irving, maybe you can just start us off with a brief overview of Western Digital for those in the investor community that may not be as familiar with the company.
Thank you, Harlan. First of all, thank you very much for hosting us.
Absolutely.
It's a pleasure to be here. Just before I go into the brief description of Western Digital, hopefully all of you, if not the majority of you, would have seen the announcement we put out this morning where we've initiated a $2 billion buyback program that's effective immediately. That's part of our ongoing capital return policy, which has three core pillars to it. One is to reinvest into the business. Second is to continue to deliver the business, which we have started to do. On the 14th of April, we redeemed $1.8 billion of debt that we had to our 2026 senior notes. And then beyond that is to return 100% of access cash flow once we hit net leverage of 1 to 1.5x.
We started that process as well with the dividend that we announced at our earnings two weeks ago, which is $0.40 per share for the year. Most recently, that buyback. To briefly describe Western Digital post-21st of February with the spinoff of SanDisk, today we are a 100% hard drive-focused data storage company that is very much a proxy for the cloud. 87% of our business at the end of our last quarter, fiscal Q3, was tied to the cloud. Very much at the forefront of what's happening in cloud and in AI in the data storage space.
It's only been a few months, Irving, since you took over as Chief Executive Officer for the team. Maybe perhaps help us understand what are your top strategic priorities for the company?
Yeah, I would say there are three key priorities. First and foremost is to really get a much deeper engagement with our customers. As I mentioned, 87% of our business today is tied to the cloud. We're really focused on how we engage with the five, six largest customers that we have that make up the majority of our revenue. We have done a whole bunch of things both internally to realign predominantly the entire organization across sales, product management, engineering, customer technical support, all the way down to supply chain to have dedicated teams for each of these large customers so that we are able to better understand their needs today and how we support that with our roadmap, but also to understand where their architectural shifts are going forward, especially as we move into this world of AI as well.
We're able to bring the full power of Western Digital to the fore both in terms of the short, medium, and long term as well. That's one key pillar. Second is around continuing to deliver leading-edge products and continuing to deliver IP innovations to our customers. That's really on making sure we continue to deliver on our leading-edge PMR technology transitioning to Hammer in the not too distant future, but also leveraging the tremendous amount of intellectual property that we have in the areas of magnetics, material science, servo mechanical components as well to see how we can incubate new businesses for the future as well.
We also have a small but rapidly growing systems and platforms business that is really seeing a lot of take-up from these new AI data center environments where the scale and speed of data flows is actually really important to them. The third one is really ensuring that we continue to deliver very strong operational execution and discipline. That is something that I think we have done a great job over the last three years, really driving our yields up. In terms of when I started in the company three years ago, our yields in our hard drive portfolio were roughly about 60%. Today, our yields are closer to the 90% mark. That is really a big driver in terms of capacities that we can put out and in terms of profitability that we can deliver as well.
Obviously, making sure that we continue to be sort of relentless in terms of our cost discipline as well. If we're able to execute on those three things, being very customer-centric, delivering innovation, and being very operation efficient, that will result in sort of the financial models that we've laid out in our recent investor day as well.
Perfect. Let's start off with some of the near-term dynamics in the market. Obviously, a lot of tariff and trade-related dynamics out there. On the demand side, you're entering a period of a little bit of uncertainty. The team has identified two opposing forces affecting the business: positive demand trends from cloud and hyperscalers on strong CapEx spending intentions this year. On the flip side, the potential for indirect and maybe direct negative impacts of trade and tariffs. We will discuss the strong cloud spending trends in a moment. On trade and tariffs, talk about the potential for direct impacts to the business. Also discuss which segments of your business are more prone to tariff-driven demand-related pullbacks.
I believe on the earnings call, you did talk about the potential for some weakness within your enterprise client channel and retail parts of your business, which, though, is only about 15% of your overall business by our estimates. Maybe just discuss some of the puts and takes around trade and tariffs.
Yeah, I think that's a fair comment. Maybe to start with the demand uncertainty that you called out. We did highlight that at our recent earnings call where we highlighted the potentials of demand uncertainty that we saw in our retail distribution business and some of the enterprise OEM business, more related to not so much the tariffs directly, but more just the implication of the tariffs on the general macro environment where there's more conservatism in terms of demand spending. That's one thing that we see, and that's why we called it out. The guidance that we gave for Q4 that we're in right now really factored all those puts and takes into it. To your point on the cloud, we see the cloud continues to be very robust.
We revalidated with all our large hyperscale customers not long after the reciprocal tariffs were announced whether there was any change in their demand profiles. The very short answer was no. In fact, we've moved the whole model to having what we call long-term agreements, 12-month agreements to our customers. We already have started to receive orders from our customers, two of our largest customers, all the way up to the first half of calendar year 2026 already. Related to the tariffs directly, it's a very fluid and volatile environment. Yesterday, I think we had some positive news on it overall. The first point to note is we don't ship anything from China to the U.S., so we're not subject to those tariffs. We have a very broad supply chain that involves the U.S. and different parts of Asia.
We have teams today that are working very closely with our customers to make sure we avoid any supply disruptions and also look at ways in which we continue to create tariff mitigation strategies for them. Longer term, obviously, we're looking more at where sort of supply chain shifts might be happening with our customers and how we need to support it so we can continue to remain agile and that our supply chains really mirror what our customer supply chains are going to look like. I would say if you summarize that up in the near term and unless things change materially, we do not see any major impact on the tariffs just yet.
As you mentioned, 75%-85% of your business is cloud-focused, right, with your leadership in high-capacity nearline hard drives. Cloud CapEx spending environment continues strong this year, right? The top four cloud and hyperscalers, 40% plus growth in CapEx, primarily driven by AI and should drive another strong year for the team, as you mentioned, for nearline HDD. That being said, there's been some concern around the potential for some AI compute and storage digestion, right? Has the team seen any recent changes in your forward LTA backlog suggesting a more cautious CapEx spending profile? Are the LTA discussions into next year suggesting continued cloud storage spending growth into calendar 2026?
Yeah, we see the outlook, at least for the entirety of our hyperscale business, continuing to be very robust at the end of the calendar year. As I've mentioned, we also have visibility with two of our largest customers well into the first half of calendar year 2026. We have not seen any pullback from that regards as well. It is also very important to recognize that the LTAs that we have with them have gone through quite a significant shift. In the past, we used to only get three- to six-month visibility. Today, we are getting 12-month visibility. That result has really been driven by a couple of factors. One, we and I think the industry in general has gone through quite a significant supply reset coming out of the down cycle that we had recently where we have taken a lot of capacity offline.
Secondly, as the business has structurally changed to be predominantly cloud and hyperscaler, we are talking about drives that are really very high capacity, very sophisticated, which have some components in them that have 50-week lead times. Given that tight supply-demand equation, long lead times, I think there is also a much better appreciation of our customers that the supply chain cannot flex as quickly as they would like if they do not give us that forward-looking visibility. I think we have gotten to a place where there is a much healthier supply-demand balance, there is greater appreciation for what is happening within our supply chain from our customers, and a greater partnership from them to recognize that the healthy supply chain and the good supply-demand balance is in the best interest of both parties.
Bringing back the strong cloud spending dynamics, right, back to your business of storage. Back in February at your analyst day, you outlined a three-year nearline exabyte growth figure of about 20%-25%, which is what, interestingly enough, some of the third-party research firms are forecasting for this calendar year, right, about 23% exabyte growth, which is consistent with the strong data center CapEx spending trends this year. Given your LTAs, strong visibility, does your forward demand pipeline profile also suggest a kind of 22%-24% exabyte shipment growth profile this calendar year or better for the WD Team?
Yeah, I think the trend that you laid out, we shared at our investor day, a 23% CAGR, right, over the calendar year 2024 to 2028 timeframe. I think that's the trajectory that we're tracking at. Within that 23% growth, we think 15% CAGR is tied to just the organic growth of the cloud with an additional 8% kicker that's AI-related. I would say that AI component today is still relatively nascent. Why I say that is actually it's there, but it's not something we can isolate specifically to an AI-specific application because what we're seeing is actually customers are deploying a lot of AI use cases into their core services. If you take a social media company, the ability to use AI to generate much more customized targeting to an individual, they are seeing the benefit of that.
They are seeing the better yields that they're getting from their programs. They are investing more into AI specifically as well. Another great example is, as we look at AI, if you use an autonomous vehicle example where autonomous vehicles are driving around, taking images of the surroundings, it's not economical for you to continue to drive a vehicle around the road. They are using those images and using AI to create synthetic data to train those vehicles as well. In a way, AI is generating more data that's resulting in more storage for us as well. It is a self-fulfilling virtual cycle that's really driving that growth.
If nearline exabyte growth is 22%, 23% this calendar year, that would suggest exabyte growth in the second half calendar year, about 15% year-over-year growth. The WD team, one of the characteristics of the WD team, and maybe we'll get to this a little bit later as well, has been the extreme discipline on capacity expansion. In fact, the team has taken out capacity over the past several years. As we think about 15% year-over-year growth in exabytes in calendar second half, I assume you can drive this with areal density improvements and not so much having to put in place new physical capacity. Is that a fair assumption?
I think it's a very fair assumption. We have no plans to put in place any incremental capacity from a unit standpoint. A lot of the focus is on areal density improvements. One of the innovations that we bring that we're very excited about is obviously our ultra SMR technology. That gives us a 20% capacity uplift over CMR, the standard recording medium, and an additional 10% capacity uplift on SMR technology. That's a great way through software enhancements. We're able to deliver higher capacity without putting in the appropriate CapEx that's tied to it. Any CapEx that we are putting into the business right now that we've indicated is going to be in the 4%-6% range. In our model, we're running slightly beneath that right now.
It's more focused around technology transitions and these innovations that are able to deliver areal density improvements rather than in terms of manufacturing capacity output.
Let's talk about the impact. You brought up a little bit about the impact of artificial intelligence and machine learning. Much of the strong cloud CapEx spending growth last year and this year has been to support AI-based compute workloads. For pre-training, post-training, optimization, inferencing, the data does originate and ultimately gets fed back to these large data lakes that reside on high-capacity HDD. The high IOPS portion of the workload is being supported by enterprise SSD, but you still need HDD. You did outline back in the February analyst day the requirements for purpose-built system-level storage and IO solutions for AI workloads. What is the WD team doing here to take advantage of this trend?
Yeah, so look, if you look at AI data workloads, especially AI object storage, which is what's basically used in HDDs, Harlan, you're absolutely right. They form the foundation of the data lakes that AI models are being used for training, post-training optimization, and inference as well. The benefit of using hard drives is they have very cost-effective bandwidth at the end of the day. The one limitation that hard drives do have is on IOPS, so the input-output operations per second. If you look at a large hyperscale data center environment, which these are our largest customers, we work very closely with them, the way they've architected their data centers is to have a very thin TLC layer of SSDs that really act as caches. They solve the IOPS needs.
Then behind them, there's a massive fleet of HDDs that really provide them that cost efficiency from a bandwidth standpoint. I think if you look at the recent blog by Google on their Colossus data center architecture, that gives you a very good insight in terms of what a large hyperscale production data center environment looks like. That's very reflective of that sort of thin layer of TLC to drive IOPS caching and then supported by a massive amount of cost-effective HDD in the background behind that.
I'm going to start talking about some of the product roadmaps with Irving. Before that, though, do we have any questions from the audience? If you do, feel free to raise your hand and we'll get a mic over to you. Any questions from the audience? Yeah, there's been a lot of.
Just one right here.
Hey, thanks very much for taking the question.
Pleasure.
With regard to how quickly you expect to grow supply, would you be expecting to kind of be right in the middle of your exabyte growth rate from a demand perspective, or would you try to kind of come in at the lower end to leave some buffer?
Yeah, I think, as I mentioned, our focus is really on trying to fulfill the exabyte demand growth by areal density improvement. Until we see that sustained demand above that 23% CAGR that Harlan highlighted in the very onset, we do not have any plans to put any additional CapEx into capacity enhancements.
It's pretty amazing, right? I think I was just trying to calculate it recently. We cover some of the memory supplies out in the market. We're coming out of a pretty deep memory-related downturn. I still calculate, and correct me if I'm wrong on this one, I still calculate, though, that enterprise SSD price per bit is still about six to eight times higher than nearline HDD price per bit. Is that about kind of the same range you guys are thinking about?
Yeah, that's very consistent with what we are seeing. In fact, going back to our investor day, we laid out a path where we look to maintain very consistently a 6x delta on a dollar per terabyte basis between our nearline drives versus a 128 TB enterprise SSD. If we can continue to maintain that, and we've laid out a roadmap to ensure that we can do that up to fiscal year 2030, that will translate into a 3.6x delta from a TCO standpoint. We work very closely on those numbers with one of our large hyperscale customers to validate that we had the right parameters that we're using and the right costs. That's something that we see we're laser-focused on and continuing to drive that cost delta between us.
I mean, it's pretty amazing. I mean, that delta has been pretty fixed for so many years now. To your point, it is the areal density innovations of the industry that really continues to drive that gap.
Yeah, I think the other point is just not the areal density improvement. If you look at the capacity output of the hard drive industry as a whole, for the memory semiconductor industry to match that, that's going to be a massive amount of CapEx that needs to be spent. Because I think if you add all the NAND output in a year, it more or less still doesn't meet the entire HDD industry output. And the capital intensity of flash versus HDD on a CapEx per exabyte level is almost 50 to 1. I actually think the hard drive industry, because of this transition to cloud and AI and it having transformed to become a data center-centric business, has gone through a significant renaissance.
Yeah, absolutely. On the product roadmaps, qual, execution, back in February, you talked about ramping to 1 million units per quarter in the second half of this calendar year, right, in your 11-platter, 26 TB CMR, and your 32 TB ultra SMR drives. You actually drove 800,000 shipments in the March quarter, so you outperformed your targets. You're already looking to deliver 1 million plus units this quarter, the June quarter. What is driving the better ramp adoption curve? Is it faster qual by your customers? Is it better operational execution by the WD team?
Yeah, I think it's a combination of all things. We were very pleased with the results in terms of the new 26 TB CMR, 32 TB ultra SMR portfolio that we launched. It has been great take-up by customers, as you highlighted, over 800,000 units shipped last quarter. We are well on track to ship well over a million units this quarter. The qual of that product was two quarters, depending on the customers within one to two quarters. That is part of our whole product roadmap strategy, is to give customers on a regular basis very predictable capacity increases that are reliable, that are scalable. When I mention scalable, that means we can consistently ship them well over a million drives a quarter post-qualification and with the right yields. I think that's very important for us.
The customer qualification, the RAM, that's really a function of how comfortable customers are with the technology and its reliability, the ability to scale up very quickly and to get yields to the point I mentioned in the 80%-90% range. That's really a function of the great work that our operations team has done. Those two factors coming together has been, I think, key drivers of both the top line growth and the margin expansion that we've seen as well.
On the next-gen platforms, you're targeting your next-gen EPMR-based 28 TB CMR, 36 TB ultra SMR drives for first half of next year qual. I assume you have to launch these programs in the second half of this calendar year to hit your target. Is the team still on track for this, or is this roadmap also maybe getting slightly pulled forward a bit as well?
I would say we're very much on track. Right now, again, because it's a very known technology to our customers, we expect qualification processes to go similar to what we just experienced with our current platform, one to two quarters. Again, immediately after qualification, we look to ship over a million units the quarter. After that, we're always looking to see how we can pull things forward, but we're on track for that roadmap right now.
On your future generation transition to Hammer technology, the team's been working on Hammer now for well over 10 years. You've been able to exploit your current EPMR technology to continue to drive areal density cost per bit, which does give the team more time to fully optimize your Hammer technology. Plan of record, start qual for Hammer, second half calendar 2026, ramp first half of 2027 with 40 terabyte plus capacities. You're also working closely with two of your large customers on Hammer. What's the feedback been like from customers and confidence level on executing to the timelines?
Yeah, as you said, we've been working on Hammer for quite a while. We just haven't been publicly talking about it too much. The roadmap that you highlighted was what we shared at Investor Day, but that's with the financial community. We've been sharing the roadmap with our customers for well over a year before that already. The feedback that we've gotten from them is that if we can execute to that roadmap, the capacity points, the timing, we can drive the scalability and reliability that we always do that's equal or better than our EPMR portfolio, they feel very comfortable with that timeline. We actually have engineering samples with two of our large hyperscale customers already today. We have weekly dialogue with them on how we're tracking in terms of the testing that we're doing with them. The feedback has been very consistent.
We're meeting all the milestones that we have jointly set. Any feedback we do get from them, we are incorporating that and we're delivering them new drives every quarter in terms of the engineering sample. I would say that's something we're laser-focused on doing. Not only just delivering the capacity points that you highlighted, we're looking at the high end of 44 on the ultra SMR portfolio. I want to emphasize making sure we can do that with the right reliability that's equal or better than what we have with our EPMR portfolio, making sure we can deliver that with the right yields and we can deliver the amounts of product that they need at scale as well. We're not only doing work just on the engineering side to make sure the technology can work and is reliable.
There's a lot of partnership going on between our engineering and our operations team to also make sure that even as we are working on the technology, we're working on the manufacturability of that technology to be able to deliver it at scale and the right yields.
Before I turn, I'm going to start talking about manufacturing and operational execution and financials. Did anybody have any questions for the team? How do you perceive the Hammer technology of your main competitor? Once you are live on your technology, do you see any risk of price competition or healthy environments? Maybe the third question on the CapEx, do you need additional CapEx to ramp up your Hammer technology or will you convert some current manufacturing capabilities?
Yep. Maybe I'll start with the CapEx question and I'll do it in reverse order. From a CapEx standpoint, as I've articulated, the only CapEx we're putting into is for technology transitions. That's related to both the next generation EPMR product that we have that we're going to start qualification in the first half of 2026 and for hammer as well. A lot of the investment that we have been making in the past can be leveraged for hammer as well. It's not going to be a completely greenfield investment. A lot of this is using the existing equipment we have and just provide some technology upgrades to be able to support the hammer transition. I think that's why we've guided that 4-6% range, and we will continue to be very disciplined in terms of our investments from that standpoint.
In terms of the roadmap, look, I think we laid out a roadmap that we just spent a fair amount of time covering. I think the best response I can give to that is the feedback that we've gotten from our customers. If we can execute to that timeline, their capacity points, they feel quite comfortable with our ability to meet their needs and where they see their data center platforms moving going forward. If we are able to do that, we continue to deliver the value that they can benefit from so that we can capture some of that as well. I think pricing would take care of itself as well.
At the prior analyst day back in May of 2022, the team laid out a three to five year target of 31%-34% gross margins. You fast forward three years and your new targets are greater than 38% gross margins. You're currently sitting at 40% plus gross margins, almost 500-700 basis points better gross margin performance versus your prior expectations. Even back in 2022, the mix was already 75% cloud. It seems like much of the better gross margin trajectory is actually operational efficiencies and cost downs. And coincidentally or not, I mean, you had just joined as Head of Global Operations in 2022. How much of the strong gross margin performance has actually been driven by operational execution? What on the op side has the team done to drive better efficiencies?
Yeah, I think overall the performance we delivered has been a great team effort. I think there are two factors to it. One is, as I mentioned, there's been a structural change within the business as we've really pivoted the business to be predominantly a cloud hyperscale-centric business. Our ability to deliver new innovations with higher capacity points has been an ability for us to drive better pricing. That has been one of the drivers for that gross margin expansion. The second, to your point, a really, really strong focus within on the operations side across two dimensions. One is to be really disciplined and laser-focused on cost takedowns, which we've executed really well on. Second, yields. I think that's something that's underappreciated in our industry.
When I started in the company three years ago, we were delivering our 18, 20 terabyte platforms, and the yields were in the 60% range. Today we've gotten them up to the high 80s, 90% depending on the platforms that we're producing. Those yields are an ability to deliver increased capacity without having to put in more CapEx. The challenge I always give the team is when we got from 65%- 80%, I said, that's the new floor. We worked from it as well. That's something we'll continue to be laser-focused as well. I can't emphasize enough when we look to make the transition to hammer, we want to make sure we get the right reliability, we get the right yields. From a margin standpoint, it's neutral to accretive in terms of what we do.
Strong gross and operating margin performance more recently, June quarter guidance, 27% up margins on 7% sequential revenue growth. Rough math leads me to incremental operating margin fall through of about 38-40% on incremental revenue growth, combination of gross margin expansion, leverage on the OpEx. Is that how to kind of think about the up margin expansion on further revenue growth?
Yeah, I think we're very focused on operating margin going forward as a business. I think we've put forward in investor day a model where we're looking at mid to high single digit growth from a revenue standpoint or at least keeping track of market growth to be in line with market growth. We put a floor of gross margin at 38%. We're tracking above that. As long as we continue to deliver innovations like ultra SMR that give us capacity uplifts and our customers' value without the associated cost in it, that's also going to continue to be accretive to the business along with the cost focus that we have.
On the OpEx side, we are spending a bit more on accelerating our Hammer development and new innovations as well related to performance of our drives, as I mentioned, on throughput, energy efficiency. We have had to invest a bit more to deal with some of the dyssynergies from the spin with SanDisk. I think what you describe is pretty much on track for what we are targeting towards.
On uses of capital, the team has targeted to return all excess free cash flow once you hit a 1-1.5x net leverage ratio. You initiated a dividend. It is about a 1% dividend yield. Is there a way to think about future dividend growth maybe in line with your free cash flow growth once you achieve your net leverage target ratio? How do we think about the new $2 billion buyback as well?
Yeah, I think it goes back to the three pillars that we laid out. First, we invest into the business, so we continue to do that. Second is to deliver the debt. I mentioned the $1.8 billion we paid, $1.8 billion we paid down. We've also publicly stated the 19.9% stake that we have in SanDisk. The disposition of that stake will go towards further deleveraging the balance sheet as well. We did indicate that the target is to complete that disposition within a year from our spin, which started on the 21st of February. We're well on track on that as well. The remainder, the 100% of excess cash, once we get to that one to one and a half range, we started, as we indicated in investor day, with a small dividend, as you mentioned, roughly 1%.
We look to increase that over time and then also complement that with buybacks as we started today as well.
Perfect. Irving, Kris, Ambrish, thank you very much for the participation today.
Our pleasure.
Appreciate it.
Thank you.